Re: iss: camera interface omap5

[Tony Lindgren <tony@atomide.com>]

* H. Nikolaus Schaller <hns@goldelico.com> [191004 16:27]:
> Hi Tony,
> 
> > Am 04.10.2019 um 18:11 schrieb Tony Lindgren <tony@atomide.com>:
> > 
> > * H. Nikolaus Schaller <hns@goldelico.com> [191004 07:25]:
> >> Hi Tony,
> >> is there a similar node for omap5 comparable to this for omap4:
> >> 
> >> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/arm/boot/dts/omap4.dtsi?h=v5.4-rc1#n176
> 
> (I just noticed there seems to be a missing empty line between 216 and 217)
> 
> > 
> > Not sure if it's the same interface as on omap4, but at
> > least TRM has "8.1.3.1 ISS Instance Summary" at 0x52000000.
> 
> Ok.
> 
> IMHO there was a patch where someone got it working a while ago so it is likely the same:
> 
> 	https://e2e.ti.com/support/interface/f/138/t/647460
> 
> Now I wonder how it can be updated to current target-module style.
> 
> Is it correct to do it similar to omap4 and replace
> 
> 	/* No child device binding, driver in staging */
> 
> by a child node that is compatible to "ti,omap4-iss".
> And there define a reg record like in the example on e2e?
> I.e. split into sysc registers for the target-module@52000000
> and driver specific registers in the child node?

Yeah something like that. Michael Alwright has a patch at [0]
with improvments for the iss staging driver to get it working.
Have you seen that one?

> Regarding clocks they likely have to stay with the target-module@52000000
> node, right?

I think it should be just something like this for
the target-module (on top of Michael's changes):

clocks = <&iss_clkctrl OMAP4_ISS_CLKCTRL 0>;
clock-names = "fck";

Then in the child node:

clocks = <&iss_clkctrl OMAP4_ISS_CLKCTRL 8>;
clock-names = "ctrlclk";

> And what about the dma nodes? Also keep in the target-module@52000000?

Those should be all in the child iss driver(s).

In general, it sounds like the iss is a private interconnect with
a control module for clocks etc, so doing it as multiple device
drivers probably makes sense.

Regards,

Tony

[0] https://github.com/allsey87/meta-builderbot/blob/master/recipes-kernel/linux/linux-stable-4.16/0008-omap4iss-Fix-multiple-bugs-and-use-device-tree.patch